[nanomsg] Re: ws transport

  • From: Andrew Starks <andrew.starks@xxxxxxxx>
  • To: "nanomsg@xxxxxxxxxxxxx" <nanomsg@xxxxxxxxxxxxx>
  • Date: Sat, 29 Nov 2014 16:48:44 -0600

On Saturday, November 29, 2014, Martin Sustrik <sustrik@xxxxxxxxxx> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi Andrew,
>
> > Would a reasonable path of exploration then be to ask: What is
> > missing from current ws libraries that would otherwise make them
> > work? Or: What needs to be changed in nanomsg to make extending it
> > to use ws a possibility; either with a current library or with a
> > new one?
>
> To use any existing ws library underneath nanomsg you'd need the
> following:
>
> All the operations have to be non-blocking, with delayed completion
> acknowledgement:
>
> 1. On POSIX systems the acknowledgement should be indicated via a file
> descriptor.
>
> 2. On Windows, the acknowledgement should be queued to an I/O
> completion port.
>
> Are you aware of any such library?
>
> Martin
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.11 (GNU/Linux)
>
> iQEcBAEBAgAGBQJUejucAAoJENTpVjxCNN9Y1pkH/2sC5sxZKaIppfqWeYeKk2ed
> 60YhoUOFo+8OkqIKgWm03XLpTsI/milFqxyzsEPqRuyAqH/aZbsg8ne7xigec+ay
> VRTIX/QoemSD4TzjPKQMJI4O+yEJFOmbhs0CzLmnDMJMIK3GalO/DpB0YcwpGA8+
> QhY1yt+TuK57jOPHJqs8AZQp/C1L/9fA+QRfwUw09oDMSgv5HgdTcXTaSLipiFf5
> jACvWQsJUtEmVva/BazHUZ4zExXFJgrOtDg65t6NdALQSW59ddGHTSp6h4rdcYBn
> WJoeQNC+99reWZRjd0PUjcOny47ygMkczYFBHlfsrV4HPpod3gBFlGTkYEtIGRg=
> =8M0J
> -----END PGP SIGNATURE-----
>
>  With my last post's caveat in effect:

No, and I don't even know that those are requirements. That is, scalability
and performance are important concerns, but there are architectural issues
that cloud the "true" separation of these transport technologies (avoiding
"later" here, because that appears to be part of the issue).

That's why I wonder (not assert) if a small service in the middle (a
broker?) couldn't server as a layer in between the two technologies. A
custom ws client could be integrated into the broker, and then the
destinction may loose meaning, except that connections through ws must be
marshaled through one or more of these....

I don't know if this is frustratingly facile for all or not. If it is,
please ignore.

-Andrew

Other related posts: